Table 1.
Comparison of some existing systems for HAR.
| Category | Reference | Sensor | Number of Activities | Number of Participants | Classification Technique | Best Accuracy |
|---|---|---|---|---|---|---|
| Environmental | Khan et al., 2022 [4] | vision sensor (Kinect V2) | 12 | 20 | Hybrid Deep Learning Model | 91% |
| Wang et al., 2017 [7] | commercial WiFi device | 8 | 25 | Hidden Markov Model | 96% | |
| Sundholm et al., 2014 [19] | textile pressure sensor | 10 | 7 | K-nearest Neighbor | 90% | |
| Wearable | Pirsiavash et al., 2012 [3] | GoPro camera | 18 | 20 | Support Vector Machine | 77% |
| Jamieson et al., 2021 [13] | accelerometer (ActivPAL) | 5 | 12 | Support Vector Machine and Long-Short Term Memory | 77% | |
| Altun et al., 2010 [12] | miniature inertial sensor and magnetometer | 19 | 8 | 7 kinds of classification techniques 1 | 99% | |
| Lim et al., 2021 [16] | accelerometer, gyroscope, magnetometer, object, and ambient sensor | 18 | 4 | Deep ConvLSTM | 91% | |
| Parzer et al., 2017 [20] | textile pressure sensor | 9 | 6 | Support Vector Machine | 92% |
1 Bayesian decision-making, decision tree, least-squares method, k-nearest neighbor, dynamic time warping, support vector machine, and artificial neural networks.